As the fiber footprint expands, more critical services are provided using optical access networks. Medical, academic and business services are a few examples. Standards bodies such as FSAN (Full-Service Access Network) are aware of this trend and of the resulting need to ensure a resilient, reliable and redundant network to support such critical services. FSAN notes that such issues have been largely ignored in optical access networks to date but must be addressed in the NGPON (Next Generation Passive Optical Network) standard currently under development to ensure the “resilience” of “high value services for residential and business applications.” This is important given the expected high levels of service integration and high split ratios (up to 1:256) foreseen for NGPON. Business class services typically demand “five 9's” (99.999%) of service availability, which equates to roughly 5 minutes of down time per year.
Unlike the core network which is based on dual, counter propagating rings to ensure duplication and redundancy at every node, the optical access network typically uses a branched topology, with a single feeder/distribution fiber feeding multiple (N) fiber drops through an optical splitter to form a 1:N network. Redundancy at every point in this type of Point to Multi-point (PtMP) access network topology is cost prohibitive and difficult to implement due to this branched topology. (It is estimated that fully redundant systems have 3.5 times the cost of equivalent systems with no redundancy.) Though full network redundancy is a difficult problem, redundancy of the more critical active components in the access network can be more easily addressed. The importance of addressing this issue becomes clear when one considers that PtMP systems such as GPON (Gigabit PON) or NG-PON can serve up to 256 customers per PtMP system. A critical component failure in the network hardware, for example, in the Central Office (CO)-based electronics, would interrupt service to the entire PtMP network.